The paper examines emission inventory data available for the utility industry and determines how well the data meet atmospheric modeling requirements. Utilities are important from the perspective of their contribution to total SO2 and NOx emissions, and also because data reporting requirements for utility plants allow consistent emission estimates to be made for each U.S. facility. It was found that publicly available data on utilities can be used to estimate plant and unit level emissions by month using consistent methods. Needs for data for time periods of less than a month (weekly, daily, hourly) are more difficult and costly to meet. Continous emission monitoring data hold promise for providing emissions information for time periods as short as 1 hour, but these instruments are in place at a limited number of units. Hourly emissions for non-monitored units would have to be estimated using rough approximations; e.g, seasonal hourly load profiles. If it is important for modelers to know that utility units were not operating for a day or a week during a certain period, special studies will have to be instituted to collect the information.